[GE users] Array jobs & job priorities

reuti reuti at staff.uni-marburg.de
Wed Dec 3 19:50:14 GMT 2008


Am 03.12.2008 um 20:33 schrieb Brendon Oliver:

> Hi Reuti,
>
> On Thu, 4 Dec 2008 04:53:47 am reuti wrote:
>
>>> Job type A is much more important than B.  So we use 'qsub -p
>>> 100 ...' when it
>>> is submitted (job type B gets no '-p' priority whent submitted).
>>
>> personally I would prefer -100 and 0 instead of positive priorities,
>> but
> what do you mean by:
>> if the user has the right to use priorities over 0 it's okay of
>> course.
> ?
>
> Do you there is a config setting which allows a user to have  
> priorities >0, or
> are you implying that it's just a convention (eg. that "ordinary"  
> users use
> priorities <0, while >0 are reserved for "special" use)?

Only managers and operators can use a priority >0. Conventional users  
are restricted to values <0.


>>> Now job type B can often be an extremely large long-running job.
>>> We know this
>>> in advance, so when this happens, the job gets submitted as an
>>> array job (
>>> qsub -t 1-n ... etc.).  There will frequently be more array tasks
>>> than there
>>> are available machines in the cluster.
>>>
>>> What I want / need / would like to happen: if an array job of type
>>> B is
>>> currently running on the queue (eg. there might be 4x machines in
>>> the queue
>>> cluster, tasks 1 thru 4 are executing on those 4 boxes, tasks 5
>>> thru 10 are
>>> sitting in the queue as "pending"), if a job of type A is
>>> submitted, that it
>>> gets scheduled before the next task of job type B gets scheduled.
>>> The type A
>>> job _must_ get executed before the remaining tasks for job B are  
>>> run.
>>
>> Exactly this is what I observe - hence to do. Which SGE version are
>> you using? You can switch on "report_pjob_tickets TRUE" to check the
>> computed prioritzy, but switch it off in long term as it's time
>> consuming.
>
> Our grid is built from sge-V60u7_1_TAG-src.tar.gz on a Gentoo linux  
> box.
> We're not really in a position to upgrade to a more recent version  
> either
> (we've got ~100 or so machines in the production grid running 24x7  
> so an
> upgrade isn't something to be take lightly ;-) ).

Even in 6.0u7 I see the same behavior. The one you expect.

> Where would I find this "report_pjob_tickets" setting? Is that in the
> scheduler config or in the queue?

It's in the scheduler config. Without it being set the job listing is  
just by jobnumber.

-- Reuti


> Thanks for your reply.
>
> Regards,
>
> - Brendon.
>
> -- 
>
>  06:25:04 up 39 days,  8:23,  2 users,  load average: 0.16, 0.20, 0.13
>
> ------------------------------------------------------
> http://gridengine.sunsource.net/ds/viewMessage.do? 
> dsForumId=38&dsMessageId=90940
>
> To unsubscribe from this discussion, e-mail: [users- 
> unsubscribe at gridengine.sunsource.net].

------------------------------------------------------
http://gridengine.sunsource.net/ds/viewMessage.do?dsForumId=38&dsMessageId=90941

To unsubscribe from this discussion, e-mail: [users-unsubscribe at gridengine.sunsource.net].



More information about the gridengine-users mailing list